Permafrost thaw sensitivity prediction using surficial geology, topography, and remote-sensing imagery: a data-driven neural network approach

Abstract

Seasonal or degradational thaw subsidence of permafrost terrain affects the landscape, hydrology, and sustainability of permafrost as an engineering substrate. We perform permafrost thaw sensitivity prediction via supervised classification of a feature set consisting of geological, topographic, and multi-spectral variables over continuous permafrost near Rankin Inlet, Nunavut, Canada. We build a reference classification of thaw sensitivity using process-based categorization of seasonal subsidence as measured from differential interferometric synthetic aperture radar whereby categories of thaw sensitivity are reflective of ground ice conditions. Classification is performed using a neural network trained on both dispersed and parcel-based reference data. For Low, Medium, High, and Very High thaw sensitivity categories, generalized classification accuracy is 70.8% for 20.6 km2 of dispersed training data. In all cases, the majority classes of Low and Medium thaw sensitivity are predicted with higher accuracy and more certainty, while the minority classes of High and Very High thaw sensitivity are under-predicted. Minority classes can be combined to improve accuracy at the expense of a reduced level of discrimination. The two-class problem can be classified with an accuracy of 81.8%, thereby effectively distinguishing between stable and unstable ground. The method is applicable to similar low-Arctic permafrost terrain with geological and topographical controls on thaw sensitivity. However, generalized accuracy is reduced for parcel-based training, indicating that reference samples are not totally representative for inference beyond the parcel, and any deployment of the network to other geographical regions would benefit from full or partial re-training with local data.

Landform mapping, elevation modelling, and thaw subsidence estimation for permafrost terrain using a consumer-grade remotely-piloted aircraft

“…Away from the site perimeter, propagation of the vertical errors at the GCPs (Table 1) suggests a change detection limit of approximately 3 cm (95% confidence). This level of precision is comparable to that of DInSAR and is sufficient for measuring seasonal thaw subsidence for the region (Oldenborger et al 2022). In contrast, propagation of the GCP matching errors for the resampled RPA DEM (Table 1) or elevation discrepancy (Table 2) suggests a change detection limit of approximately 20-25 cm (95% confidence).…”

“…The low-lying marine sediments host an ice-rich top of permafrost and exhibit seasonal thaw subsidence of 0-10 cm at multiple scales. Seasonal subsidence greater than approximately 3.5 cm is indicative of thawing of excess ice (Oldenborger et al 2022).…”

Quantification of active layer depth at multiple scales in Interior Alaska permafrost

“…Machine learning models can assist in predicting ALT and other related features from specific measurement locations in cold regions across a wider area (Panda et al 2012, Pastick et al 2015, Gulbrandsen et al 2016, Shi et al 2018, Whitley et al 2018, Baral and Haq 2020. To strengthen these models, inputs such as multispectral (Oldenborger et al 2022) and hyperspectral (Anderson et al 2019) remote sensing imagery, airborne geophysics (Jorgenson and Grosse 2016), terrestrial laser scanning (Anders et al 2020), light detection and ranging (LiDAR; Brown et al 2016), normalized difference vegetation index (NDVI; Beck et al 2015), and vegetation data (Heijmans et al 2022) can be utilized.…”

Mineral Element Identification in Remote Sensing Imagery: A Fusion Approach Using CH-Tucker Decomposition and RFDNet

“…In recent times, remote sensing technology has emerged as a cornerstone in geological exploration and mineral prospecting, proving itself instrumental in acquiring terrestrial surface information [1][2][3][4]. Images procured through remote sensing offer a wealth of geological information, enabling both macroscopic and microscopic viewpoints of mineral distribution.…”

“…Such images are predominantly sourced from various remote sensing satellites and aircraft, encapsulating vast terrestrial expanses. Recognized for their consecutive, expansive coverage, and periodic nature, these images are pivotal for the identification of mineral elements [1][2][3][4].…”